The present invention relates generally to atomic or molecular frequency standard masers and more particularly to such a maser that includes a variable volume, constant surface area storage bulb.
Masers employing collimated beams of atomic or molecular particles have been developed to provide frequency standards having very great stability. Typically, the beam consists of hydrogen atoms having a predetermined energy level, which atoms are confined for long periods in a storage bulb, usually having a Teflon coating. The confined masing atoms stimulate an electromagnetic field in a resonant cavity that is electromagnetically coupled to the storage bulb. Because the masing atoms are confined in the storage bulb, the stimulated frequency of the hydrogen maser is changed. Collisions of the hydrogen atoms with walls of the storage bulb cause this frequency change, known as the wall shift, which is given by: EQU f = .phi.v/2.pi. .multidot. A/4V (1)
where:
.phi. = THE AVERAGE PHASE SHIFT OF THE MASER FOR EACH COLLISION BETWEEN THE HYDROGEN ATOMS AND THE BULB WALL;
V = THE AVERAGE SPEED OF HYDROGEN ATOMS COLLIDING WITH THE BULB WALL;
V = the volume of the bulb; and
A = the surface area of the bulb wall.
The expression 4V/A, which is designated as .lambda., represents the mean free path of hydrogen atoms in the bulb between collisions.
In order to use the hydrogen maser as a primary frequency standard, it is necessary to correct for the wall shift. One technique which has been employed to correct for wall shift involves varying the length of the mean free path between wall collisions by operating different hydrogen masers with differing size storage bulbs and to make frequency comparisons between outputs of the masers. This technique has not produced particularly good results because the average phase shift per collision has not reproduced well from one bulb to another.
Another approach, which has proven more successful, involves providing a maser with a flexible, Teflon storage bulb so that the value of .lambda. can be varied while maintaining the average phase shift per collision constant. This approach does not require changing bulbs, yet provides for wall shift correction. In the variable volume arrangement, it is necessary to have an accurate knowledge of the ratio of the volumes of the bulbs in at least two different volumetric conditions. In addition, it is necessary to determine accurately the average phase shift per collision, as well as the corrected frequency, i.e., the frequency after the wall shift is considered. These average phase shift and frequency corrections must remain constant for the different volumes of the bulb. With careful measurement, the ratios of the two volumes can be determined to approximately 0.1%.
It has been found, however, that the volume ratio strongly affects the wall shift, and therefore, the frequency error. As the volume ratio approaches unity, the uncertainty in determining the correct frequency goes to infinity for given errors in measurements of the volume ratio and the frequencies of the maser for the different volumes. In a device that was actually constructed, the magnitude of the volume ratios was severely limited because the filling factor of the bulb is degraded at a compressed, relatively small volume of the bulb. In this device, having a bulb volume ratio in the range of 1.118:1.37, there were significant errors that could not be tolerated for frequency standards. In addition, it was found that changes in the surface properties of the Teflon bulb during the measurement process induced changes in stress of the bulb when the bulb volume was changed. The uncertainty resulting from the stress effects on the storage bulbs was a major factor in preventing the adoption of this variable volume device as a frequency standard.
To overcome the stress problem associated with the flexible bulb, it has been proposed to employ a thin, flexible Teflon cone attached to a rigid cylinder as a variable volume storage bulb. The advantage of this configuration is that a thin cone can be inverted in such a way that only the edges of the cone are stressed. By stressing only the edges of the cone, the region of possible stress is a negligibly small area, to eliminate the uncertainty due to surface stressing. However, in the device which was actually constructed, the maximum volume ratio was limited to 1.3 because of difficulties in obtaining maser oscillation with an inverted cone. The accuracy of the device was limited to 2.4 .times. 10.sup.-12, rather than to the theoretical value of 10.sup.-14, because of the small maximum volume ratio, drifts in a reference maser that was beat against the maser including the thin, flexible Teflon cone, and because areas in the cone became exposed when the cone was inverted. The bulb also was asymmetrical, making it especially susceptible to magnetic inhomogeneity shifts, which have been found to cause errors as large as a few parts in 10.sup.12.
To overcome the problems associated with the thin, flexible Teflon cone device, it was proposed to combine the flexible cone with a large storage box hydrogen maser. In this device, a flexible cone is outside of the microwave cavity so the magnitude of the volume ratio is not limited by the desired frequency of oscillation. This device had the advantage of reducing magnetic inhomogeneity, and because the device has a line width factor of 10 narrower than a conventional hydrogen maser, anomalous spin change effects are correspondingly reduced.
The large storage box device included a pair of small resonant cavities, each having a length to diameter ratio of 1. Both cavities were external to the box which had a length, in the expanded position, of five feet, and a similar diameter. One of the cavities was a low level, output cavity, while the other cavity was a high level driving cavity, responsive to the output of a +60 db gain amplifier driven by the low level cavity. The resulting, high gain feedback loop is necesssary because the vast majority of the large box, where the stimulated emission of electromagnetic energy occurs, is outside of the resonant cavity, i.e., hydrogen atoms are not subjected to the electromagnetic fields of the cavity while they are in the vast majority of the large box. However, the amplifier in the feedback loop of the maser produces phase instability in the stimulated emission; the phase instability results in frequency drift, so that the accuracy of the standard is limited to one part in 10.sup.13. Frequency instability of the large box device was also caused by a relatively large uncertainty in the value of the volume ratio for the big box in its differing sizes. This uncertainty is due, to a large extent, to the extremely large volume of the box and the volume changes which are necessary.
Another device that has been proposed is the so-called Concertina hydrogen maser wherein a variable volume storage bulb is formed of a flexible Teflon film bellows which is located entirely within the resonant cavity. The bellows stretch effects are cancelled, to a first order, because of its configuration. In addition, the device has the great advantage of allowing measurements to be made over a continuous range of calibrated volumes. However, the asymmetrical arrangement of the storage bulb in the microwave, resonant cavity makes the device very susceptible to magnetic inhomogeneity problems.
In studying hydrogen masers, it has previously been discovered that a zero frequency wall shift occurs at approximately 100.degree. C for FEP Teflon storage bulbs. Variable volume storage bulb devices have been proposed as null detectors for the zero wall shift point. The advantage of such devices is that there is no need to know the volume ratio accurately to calibrate for wall shift. The device can be operated in an automated fashion to maintain the zero wall shift. However, the problem in attaining the zero wall shift is that there is a tendency for the Teflon to outgas at the elevated temperature. Of course, outgassing is undesirable because it causes particles other than hydrogen to be produced in the bulb and interfere with the maser operation.